Compare AI models

Specifications
Parameters
14B6B
Context window
2k2k
Benchmarks
HumanEval
32.3%0%
MATH
24.8%3.1%
MMLU
66.3%25.2%
GSM8K
61.3%1.5%
Overview
CompanyQwenZ.ai
Release dateSep 25 2023Mar 14 2023
AccessOpen WeightOpen Weight
Model detailsView modelView model

Frequently asked questions

Qwen-14B leads ChatGLM-6B on 4 of the 4 benchmarks they both report (HumanEval, MATH, MMLU, GSM8K). ChatGLM-6B shipped 195 days before Qwen-14B, so benchmark comparisons should account for the intervening progress.

Qwen-14B has 14B parameters, while ChatGLM-6B has 6B. Context windows are 2k (Qwen-14B) vs 2k (ChatGLM-6B).

On HumanEval, Qwen-14B leads at 32.3% vs ChatGLM-6B at 0%. On MATH, Qwen-14B leads at 24.8% vs ChatGLM-6B at 3.1%. On MMLU, Qwen-14B leads at 66.3% vs ChatGLM-6B at 25.2%. On GSM8K, Qwen-14B leads at 61.3% vs ChatGLM-6B at 1.5%.